REAL TIME FACE MASK DETECTION USING MOBILENET WITH HARR CASCADED TECHNIQUE
NITHYAA SHRI N S1, PADMAJA T2, UMARANI C3
1Department of CSE-UG, Kingston Engineering College, Vellore-59
2Department of CSE-UG, Kingston Engineering College, Vellore-59
3Assistant professor, Department of CSE, Kingston Engineering College, Vellore-59
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Wearing a mask in most of the non-pharmaceutical measures that may be used to reduce the main source of COVID droplets expelled via way of means of an infected individual. To make contributions towards the communal health, this paper pursuits to devise a highly accurate and real-time method that could effectively stumble on non-masks faces in public and thus, implement them to put on mask. Although several researchers have dedicated efforts in designing efficient algorithms for face detection and recognition, there exists an important distinction between ‘detection of the face beneath masks’ and ‘detection of masks over the face. Manual real-time tracking of facemask sporting for a huge group of human beings is turning into a hard task. The aim of this paper is to apply deep learning (DL), which has proven incredible outcomes in lots of real-existence applications, to make sure efficient real-time facemask detection. The proposed method is primarily based on two steps. An off-line step aiming to create a DL model this is capable of stumble on and discover facemasks and whether they are appropriately worn or not. An on-line step that deploys the DL version at edge computing on the way to stumble on mask in real-time. In this study, we advocate to apply MobileNetV2 to detect facemask in real-time. Several experiments are performed and display true performances of the proposed method (99% for training and testing accuracy).
Key Words: Face mask detection, Deep learning.
